iSDoH

iSDoH is the abbreviation of Intelligent Social and Environmental Determinants of Health

Demo

Background and Introduction

The SDoH (Social Determints of Health) and environmental datasets are getting more and more popular for clinical research and highly needed by researchers around the world. However, because the various datasets are collected and shared by different government organizations, research institutes and so on, there is lack of a combined database for researchers to check, visualize and download data in a quick and reliable manner.

So we started the journey and tried to provide a solution for below areas:

1. A GIS database
2. Quick data ingestion with limited manual jobs
3. A slim web application for data catalog, visualization and even analysis
4. A local running toolkits for geocoding

Database design

We seletced PostgreSQL database and GIS extension to fulfill our requests. Inspired by OHDSI GIS working group, we designed our database with the most fundmental components and tables. As SDoH and environmental datasets are highly involved with geo-information, we ingested the Census Tract, ZCTA, County and States shapefiles into the database and link the variable tables with the geo-tables.

• data_source table is used to store significant metadata from all source datasets includes SDoH datasets, environmental data and geomentry data.

[placeholder of a metadata table]

• For each SDoH and environmental data, there are two tables generated

  • variable_index table, which is applied to record the important attributes /properties of each variable
  • variable_value table, which is applied to store the values of the variable for each geocoded locations

• For each geometry data, we stored them as a single table to link with variables through geoCodes such as FIPS, ZCTA, COUNTY and STATE.

Toolkits

Since the most SDoH datasets have the Census Tract(FIPS) as boundary type, the toolkits are developed for investigators to geocode the source address to coordinations or FIPS codes.

To maintain the privacy of patient's address, investigators need to execute the geocoding toolkits locally with the help of DeGauss. We prepared the scripts for both file processing and fetch data directly from OMOP databases.

If your institute already provides the geocoding service, please ignore the toolkit.

Detailed execution instructions are recorded at here

Web Application

We created a web application to provide the investigators with the access to :

• data catalog
• SDoH and Env data linkage tool
• data visuallization tool

In the next release, we are planning to provide a basic data analyzer tool.

The web application is deployed at HiPerGator PubApps